financial-intel

When To Use

Use this skill for a fixed five-call crypto market-research bundle when the user wants several independent signals in one brief:

• current token spot price;
• token market, volume, supply, and price-change metrics;
• one user-selected US equity or ETF benchmark quote;
• chain-wide aggregated smart-money holdings; and
• current news and market context with source URLs.

The user must identify one coherent crypto asset across symbol, coin, assetChain, and query, and explicitly select the ticker benchmark. For example, ETH / ethereum / ethereum / QQQ / an Ethereum-specific current news query is coherent. Do not silently combine an asset with unrelated manifest defaults.

This is a research workflow, not personalized financial advice. It does not place trades, predict returns, or decide whether the user should buy, hold, or sell. For an equity-only study or a single price/news lookup, choose a smaller skill or free endpoint discovery instead of paying for all five calls.

Workflow

1. Install the vetted recipe:

   selat skill install financial-intel

2. Collect and validate all five required inputs:

   • symbol: token symbol, such as ETH;
   • coin: matching CoinGecko ID, such as ethereum;
   • ticker: the explicit US equity or ETF benchmark, such as QQQ;
   • assetChain: Nansen data chain, such as ethereum; and
   • query: current, asset-specific search wording.

   Confirm the symbol and CoinGecko ID refer to the same asset. Explain that the Nansen step describes the whole selected chain, not that asset alone.

3. Probe all five payment challenges for free:

   selat skill verify ~/.config/selat/skills/financial-intel \
     --symbol ETH \
     --coin ethereum \
     --ticker QQQ \
     --assetChain ethereum \
     --query "Ethereum ETH ETF flows regulation latest news" \
     --live-probe

4. Show every live quote, the expected cumulative total, and the sum of all per-step caps. Obtain explicit approval for the exact workload and a proposed cumulative session budget. Verification without --pay is free.

5. Only after approval and a spendable Gateway balance, arm the approved budget, run the fixed bundle once, and disarm the budget after success or failure:

   selat budget start --amount <approved-cumulative-cap>
   selat skill run financial-intel \
     --symbol ETH \
     --coin ethereum \
     --ticker QQQ \
     --assetChain ethereum \
     --query "Ethereum ETH ETF flows regulation latest news"
   selat budget stop

The CLI runs every manifest step in order and continues after an individual failure. Inspect each result and payment history; a final partial failure does not mean earlier calls were uncharged. Never retry a paid failure until its transaction state is checked, prices are re-probed, and the retry is separately approved.

Fixed Steps

1. Alchemy spot price — GET /prices/v1/tokens/by-symbol for symbol.
2. CoinGecko token market data — POST /coingecko/coins-markets for coin, bounded to one result with 1h/24h/7d changes and sparkline data.
3. Alpha Vantage benchmark quote — POST /alphavantage/global-quote for the user-selected ticker. Treat correlation or causal interpretations as hypotheses, not facts established by one quote.
4. Nansen chain-level smart-money holdings — POST /api/v1/smart-money/holdings for assetChain. The response aggregates token holdings by sophisticated traders and funds across that chain; it is not filtered to symbol or coin.
5. Exa current context — POST /search for the explicit query, returning eight results with bounded text and URLs.

Inputs And Outputs

Each step returns provider JSON plus the CLI's per-step status. Produce the final brief in this order:

1. retrieval time and exact identifiers used;
2. spot and token-market snapshot;
3. benchmark quote, clearly labelled as context;
4. chain-level smart-money observations, without asset-level overclaiming;
5. dated news claims with source URLs;
6. confirmations, contradictions, missing data, and limitations; and
7. neutral scenarios and risks, not a personal buy/sell recommendation.

Keep conflicting values and their timestamps visible rather than averaging them or silently selecting one provider.

Rails And Costs

The manifest declares two payment-protocol categories and the 2026-10-04 live verification observes two routed modes:

• Alchemy and Exa: x402 via Circle Gateway, currently routed-x402. (Exa previously quoted routed-mpp; it now resolves x402.)
• CoinGecko, Alpha Vantage, and Nansen: MPP on Tempo, currently routed-mpp. Nansen is dual-protocol: its MPP challenge appears only behind the probe's Authorization: Payment retry, and the router has been observed settling Nansen over x402. The rail label follows the probe mode; check selat history for the settled rail.

All five calls currently require a reachable SELAT_ROUTER_URL. Do not infer direct versus routed execution from the provider hostname; the free live probe is authoritative.

maxAmount values are per-call ceilings, not price estimates and not a pooled run cap. Current per-step caps are $0.0015, $0.09, $0.012, $0.07, and $0.01 (~1.3–1.5x each live quote), totaling $0.1835. The manifest's top-level $0.09 is only a fallback for a step without an override — it is not a full-run cap, and every current step has an override. A separately armed session budget supplies the cumulative limit.

The free live probe on 2026-10-04 quoted $0.001, $0.063, $0.0084, $0.0525, and $0.007, for an expected total of $0.1319. Re-probe before every paid run because live prices and modes can change.

Gotchas

• Fixed pipeline, not cheapest-first. selat skill run executes all five entries and cannot stop early or select a subset.
• assetChain is a data input. Do not rename it to chain: --chain is a reserved SELAT/selat-pay settlement flag and will not reliably populate the Nansen body.
• Nansen is chain-wide. Its holdings schema accepts chains but no asset ID in this operation. Do not call its response “ETH smart-money positioning” merely because assetChain=ethereum.
• No inter-step dataflow. One provider's output does not correct or populate later inputs. Validate all identifiers before quoting or paying.
• GET versus POST. Alchemy inputs are in the query string; all other inputs are JSON request-body fields.
• A paid error may still charge. The runner continues after failures. Check transaction history before considering a retry.
• Research-only safety. Do not turn sparse, stale, or conflicting data into personalized investment advice or a deterministic price forecast.

Validation

• Static: selat skill validate ./skills/financial-intel
• Live gate, free: selat skill verify ./skills/financial-intel --symbol ETH --coin ethereum --ticker QQQ --assetChain ethereum --query "Ethereum ETH ETF flows regulation latest news" --live-probe
• Paid verification: only after fresh quotes, explicit approval, and an armed cumulative session budget, add --pay; each paid call can settle independently.

• Single-step probe, free:

  --chain base in the probe commands below is only the flag selat-pay requires today — a probe reads a free, chain-independent quote and never settles. A real paid run resolves the settlement chain from your funded Circle Gateway balance, not the manifest.

  selat-pay GET "https://x402.alchemy.com/prices/v1/tokens/by-symbol?symbols=ETH" --chain base --max-amount 0.0015 --probe-only --live-probe selat-pay POST "https://api.nansen.ai/api/v1/smart-money/holdings" --body '{"chains":["ethereum"]}' --chain base --max-amount 0.07 --probe-only --live-probe

References

• manifest.json — machine-readable fixed payment recipe.
• references/endpoints.md — endpoint schemas, routing observations, prices, and interpretation limits.
• references/agent-skill-authoring-sop.md — authoring standard.
• selat-pay — https://github.com/SELAT-AI/selat-pay

Provider and product names are used only to identify third-party services. This skill is not investment advice and does not imply provider endorsement.